IdentigonAnonymise a database, credibly.

Deterministic pseudonymisation and privacy-preserving database cloning — for test, staging, and demo environments that need real-looking data without real PII.

View on GitHub

Getting Started

Fail-closed by default

Every column must resolve to an explicit role before a run starts. Nothing is ever silently copied through unclassified, inference only ever suggests.

Deterministic & reproducible

The same source value and salt always fabricate to the same fictional value — no randomness, no drift between runs, byte-for-byte repeatable when you need it to be.

Guaranteed-fictional output

Fabricated values land in officially reserved or structurally-impossible ranges wherever one exists, so the result doesn't just look fake — it fails a real lookup.

Built-in accountability

Every run emits a DPIA report (HTML, JSON, and Markdown) recording what changed and why, so a classification never has to be trusted blind.